The homeodomain family of transcription factors plays a fundamental role in a diverse set of functions that include body plan specification, pattern formation, and cell fate determination during metazoan development. Members of this family are characterized by a helix-turn-helix DNA-binding motif known as the homeodomain. Homeodomain proteins regulate various cellular processes by specifically binding to the transcriptional control region of a target gene. These proteins have been conserved across a diverse range of species, from yeast to human. A number of inherited human disorders are caused by mutations in homeodomain-containing proteins. For example, the retroviral oncogene qin, homolog of mammalian brain factor 1 (FOXG1B), belongs to the family of winged helix transcription factors. Oncogenic transformation by Qin requires sequence-specific DNA binding. Missense mutations in the forkhead domain of Qin modulates its oncogenic transforming ability in chicken embryonic fibroblasts. We used homology model building (threading) techniques to examine the tertiary structure of wild-type c-Qin and c-Qin mutants, using the solution structure of the forkhead domain of the adipocyte-transcription factor as a template . The atomic structures of the Qin forkhead domain were generated through molecular modeling techniques, based on the alignment obtained from threading experiments. Energy calculations indicate that the Qin forkhead structure is stabilized primarily by hydrophobic interactions between residues at the helical interface. None of the missense mutations analyzed here were responsible for maintaining the most critical pairwise interactions holding the forkhead domain together. The mutated proteins form the overall structure of the forkhead domain, but the mutations do interfere with DNA binding. Bioinformatic techniques were also used to examine a functional assignment to a protein called RAG1, catalogued as a homeodomain protein in OMIM. A more in-depth bioinformatic analysis shows this assignment to be incorrect. The known biochemical functions of RAG1 are as an integrase and recombinase , functions that are not consistent with those of other homeodomain proteins .While one of the goals of functional annotation is to catalog information that would be of value in guiding experimental design and analysis, this is a textbook case where sequence similarity was detected reliably, but the functional annotation found in the public databases were incorrect. Finally, the workgroup continues to oversee the curation of the Homeodomain resource, a searchable collection of information for the homeodomain protein family. The resource is organized in a compact form and provides user-friendly interfaces for both querying the component databases and assembling customized datasets. The current release (version 5.0, October 2002) contains 1056 full-length homeodomain-containing sequences, 37 experimentally-derived structures, 81 homeodomain interactions, 84 homeodomain DNA- binding sites, and 114 homeodomain proteins implicated in human genetic disorders. A new feature of this new release is the inclusion of experimentally-derived protein-protein interaction data for homeodomain family members. All entries are cross-linked for easy retrieval of the original records from source databases. The Homeodomain Resource is freely available through the World Wide Web at http://research.nhgri.nih.gov/ homeodomain/.